Gas sensor

The present application claims priority on the basis of the Japanese Patent Application No. 2018-126301 filed on Jul. 2, 2018, and the Japanese Patent Application No. 2019-059954 filed on Mar. 27, 2019, the entire contents of which are incorporated herein by reference.

The present invention relates to a gas sensor.

In the related art, a gas sensor is known for detecting a specific gas concentration, such as NOx, in a measurement-object gas, such as automobile exhaust gases. For example, Patent Literature 1 describes a gas sensor including a layered body that includes a plurality of oxygen-ion-conductive solid electrolyte layers and electrodes provided on the solid electrolyte layers. When the concentration of NOx is to be detected by using this gas sensor, first, oxygen is pumped to the inside or outside by a measurement-object gas flow section, which is within the sensor element, and a portion outside of the sensor element, thereby adjusting the oxygen concentration of the measurement-object gas flow section. Subsequently, NOx in the measurement-object gas, which is a gas after the oxygen concentration is adjusted, is reduced, and the concentration of NOx in the measurement-object gas is detected based on a current that flows at an electrode (measurement electrode) within the sensor element in accordance with the oxygen concentration after the reduction. Furthermore, Patent Literature 2 describes a gas sensor for detecting the concentration of ammonia present in a measurement-object gas. The gas sensor detects the concentration of ammonia as follows. Ammonia is converted to NOx by being oxidized with oxygen present in the measurement-object gas, and the concentration of NOx derived from the ammonia is detected by using a method similar to that of Patent Literature 1.

PTL 1: JP 2014-190940 A

PTL 2: JP 2011-039041 A

To date, not many studies have been conducted on using, as a measurement-object gas, a low-oxygen atmosphere gas (including cases in which the measurement-object gas is a rich atmosphere gas containing unburned fuel). Recently, the present inventors measured a specific gas concentration in a measurement-object gas that was in a state of a low-oxygen atmosphere and found that a measurement accuracy decreased.

The present invention has been made to solve such a problem, and a principal object of the present invention is to suppress a specific gas concentration measurement accuracy from decreasing in a case where the measurement-object gas is a low-oxygen atmosphere.

To achieve the above-described principal object, the following configuration is employed in the present invention.

The gas sensor of the present invention includes an element body including an oxygen-ion-conductive solid electrolyte layer, a measurement-object gas flow section being provided within the element body to allow a measurement-object gas to be introduced into the measurement-object gas flow section and flow through the measurement-object gas flow section; an adjustment pump cell that adjusts an oxygen concentration of an oxygen concentration adjustment chamber, the oxygen concentration adjustment chamber being provided in the measurement-object gas flow section; a preliminary pump cell that pumps oxygen into a preliminary chamber to prevent the measurement-object gas from reaching the oxygen concentration adjustment chamber in a state in which the measurement-object gas is a low-oxygen atmosphere, the preliminary chamber being provided upstream of the oxygen concentration adjustment chamber in the measurement-object gas flow section; a measurement electrode disposed on an inner peripheral surface of a measurement chamber, the measurement chamber being provided downstream of the oxygen concentration adjustment chamber in the measurement-object gas flow section; a reference electrode that is disposed within the element body and to which a reference gas is to be introduced, the reference gas serving as a reference for detecting a specific gas concentration in the measurement-object gas; a measurement voltage detection device that detects a measurement voltage present between the reference electrode and the measurement electrode; and a specific gas concentration detection device that obtains, based on the measurement voltage, a detection value according to oxygen produced in the measurement chamber and, based on the detection value, detects the specific gas concentration in the measurement-object gas, the oxygen being oxygen derived from the specific gas.

In this gas sensor, a measurement-object gas is introduced into the measurement-object gas flow section, and then the oxygen concentration of the measurement-object gas is adjusted by the adjustment pump cell in the oxygen concentration adjustment chamber, and, after the adjustment, the measurement-object gas reaches the measurement chamber. Further, the gas sensor obtains, based on the measurement voltage, a detection value that corresponds to oxygen produced in the measurement chamber, the oxygen being oxygen derived from the specific gas, and based on the obtained detection value, the gas sensor detects the specific gas concentration in the measurement-object gas. Furthermore, to prevent a measurement-object gas from reaching the oxygen concentration adjustment chamber in a state in which the measurement-object gas is a low-oxygen atmosphere, the preliminary pump cell pumps oxygen into the preliminary chamber provided upstream of the oxygen concentration adjustment chamber. In the gas sensor of the present invention, since the preliminary pump cell supplies oxygen to a measurement-object gas prior to adjustment of the oxygen concentration as just described, it is unlikely that the measurement-object gas will be introduced into the oxygen concentration adjustment chamber in a state in which the measurement-object gas is a low-oxygen atmosphere, even in a case where the measurement-object gas is a low-oxygen atmosphere before being introduced into the measurement-object gas flow section. Hence, a decrease in measurement accuracy that occurs in a case where the measurement-object gas is a low-oxygen atmosphere is suppressed.

It is to be noted that, in the case where the specific gas is an oxide, the phrase “oxygen produced in the measurement chamber, the oxygen being oxygen derived from the specific gas” may refer to oxygen produced when the specific gas itself is reduced in the measurement chamber. In the case where the specific gas is a non-oxide gas, the phrase “oxygen produced in the measurement chamber, the oxygen being oxygen derived from the specific gas” may refer to oxygen produced when the specific gas is converted into an oxide and the resulting gas is reduced in the measurement chamber. Furthermore, the specific gas concentration detection device may obtain the detection value as follows. Based on the measurement voltage, oxygen produced in the measurement chamber may be pumped from the measurement chamber to the outside, the oxygen being oxygen derived from the specific gas, so that the oxygen concentration in the measurement chamber can reach a predetermined low concentration. When the pumping is performed, a measurement pump current flows. The measurement pump current may be the detection value. The element body may be a layered body including a plurality of stacked oxygen-ion-conductive solid electrolyte layers.

The gas sensor of the present invention may further include a preliminary pump control device. The preliminary pump control device may control the preliminary pump cell in a manner such that a constant preliminary pump current flows through the preliminary pump cell. With this configuration, oxygen can be supplied, by performing a relatively simple control, to a measurement-object gas that is a low-oxygen atmosphere in the preliminary chamber.

The gas sensor of the present invention may further include a storage device. The storage device may store information related to a relationship formula representing a relationship between the detection value and the specific gas concentration. Regardless of whether a measurement-object gas that is outside of the element body is a low-oxygen atmosphere, the specific gas concentration detection device may detect the specific gas concentration by using the relationship formula stored in the storage device, the relationship formula being a common formula. In this manner, the gas sensor of the present invention can detect the specific gas concentration accurately without using different relationship formulas for the case in which the measurement-object gas is a low-oxygen atmosphere and for the case in which the measurement-object gas is not a low-oxygen atmosphere. Hence, the gas sensor can detect the specific gas concentration readily and accurately.

In the gas sensor of the present invention, the specific gas concentration detection device detects the specific gas concentration, and the specific gas concentration may be a concentration corrected based on an oxygen concentration of the measurement-object gas that is outside of the element body. It is to be noted that, even in the case where the actual specific gas concentration (real concentration) in a measurement-object gas is uniform, the detection value may change with the oxygen concentration of a measurement-object gas that is outside of the element body, and in this case, the specific gas concentration measured based on the detection value also changes. Accordingly, by detecting the specific gas concentration by involving the oxygen-concentration-based correction, a specific gas concentration measurement accuracy is improved. The phrase “detect the specific gas concentration, and the specific gas concentration is a concentration corrected based on an oxygen concentration of the measurement-object gas” encompasses the following: cases in which the specific gas concentration is detected based on a detection value obtained after the oxygen-concentration-based correction; and cases in which, when detecting the specific gas concentration based on the detection value, the oxygen-concentration-based correction is performed, and the corrected specific gas concentration is detected.

In this case, the gas sensor of the present invention may further include a preliminary pump control device and an oxygen concentration detection device. The preliminary pump control device may control the preliminary pump cell in a manner such that a constant preliminary pump current flows through the preliminary pump cell. The oxygen concentration detection device may detect the oxygen concentration of the measurement-object gas that is outside of the element body, the oxygen concentration being detected based on the constant preliminary pump current, an adjustment pump current that flows when the adjustment pump cell pumps oxygen from the oxygen concentration adjustment chamber in a manner such that the oxygen concentration of the oxygen concentration adjustment chamber reaches a target concentration, and the target concentration. The specific gas concentration detection device may correct the specific gas concentration by using the oxygen concentration detected by the oxygen concentration detection device. It is to be noted that the constant preliminary pump current that flows through the preliminary pump cell corresponds to the flow rate of oxygen pumped into the measurement-object gas flow section by the preliminary pump cell. Furthermore, the adjustment pump current corresponds to the flow rate of oxygen pumped from the oxygen concentration adjustment chamber. Hence, the oxygen concentration of a measurement-object gas outside of the element body can be detected based on the currents and the target concentration. That is, the oxygen concentration necessary for correction can be detected by the gas sensor of the present invention.

The gas sensor of the present invention may further include a measurement-object gas-side electrode disposed at a portion that is to be exposed to the measurement-object gas that is outside of the element body. The preliminary pump cell may pump oxygen into the preliminary chamber from a vicinity of the measurement-object gas-side electrode. With this configuration, the following is possible. In comparison with, for example, a case in which oxygen is pumped into the preliminary chamber from a vicinity of the reference electrode, a decrease in measurement accuracy that may occur when the potential of the reference electrode changes as a result of a voltage drop due to the current during pumping is suppressed.

In the gas sensor of the present invention, the measurement-object gas may be an exhaust gas from an internal combustion engine, the reference gas may be air, and the preliminary pump cell may pump oxygen into the preliminary chamber from a vicinity of the reference electrode. With this configuration, the following is possible. In comparison with, for example, a case in which oxygen is pumped to the inside from exhaust gases that are outside of the element body, oxygen can be pumped into the preliminary chamber at a low applied voltage because air has a higher oxygen concentration than exhaust gases.

1 FIG. 2 FIG. 100 90 100 100 100 100 101 15 21 41 50 80 83 90 101 15 21 41 50 80 83 101 90 100 An embodiment of the present invention will now be described with reference to the drawings.is a schematic cross-sectional view schematically illustrating an example of a configuration of a gas sensor, which is an embodiment of the present invention.is a block diagram illustrating an electrical connection relationship between a controllerand individual cells. The gas sensoris attached to, for example, a pipe such as an exhaust gas pipe of an internal combustion engine, examples of which include gasoline engines and diesel engines. The gas sensordetects the specific gas concentration in a measurement-object gas, which is exhaust gases from an internal combustion engine. Examples of the specific gas include NOx and ammonia. In the present embodiment, the specific gas concentration measured by the gas sensoris the concentration of NOx. The gas sensorincludes a sensor element, cells,,,, andto, and a controller. The sensor elementhas an elongated parallelepiped shape. Each of the cells,,,, andtoincludes a portion of the sensor element. The controllercontrols the gas sensoras a whole.

101 1 2 3 4 5 6 101 2 The sensor elementis an element including a layered body in which six layers, namely a first substrate layer, a second substrate layer, a third substrate layer, a first solid electrolyte layer, a spacer layer, and a second solid electrolyte layer, are layered in this order from the bottom side, as viewed in the drawing. Each of the six layers is formed of an oxygen-ion-conductive solid electrolyte layer containing, for example, zirconia (ZrO). Furthermore, the solid electrolyte forming each of the six layers is dense and gas-tight. The sensor elementis produced as follows, for example. Ceramic green sheets corresponding to the respective layers are subjected to, for example, a predetermined process and circuit pattern printing. The resulting sheets are then layered together and subjected to firing to be unified.

1 FIG. 101 10 11 12 13 20 30 40 60 61 6 4 On the front side (left side of) of the sensor element, a gas inlet port, a first diffusion-rate-limiting portion, a buffer space, a second diffusion-rate-limiting portion, a first internal space, a third diffusion-rate-limiting portion, a second internal space, a fourth diffusion-rate-limiting portion, and a third internal spaceare formed adjacent to one another in such a manner as to be in communication with one another in this order, between the lower surface of the second solid electrolyte layerand the upper surface of the first solid electrolyte layer.

10 12 20 40 61 101 5 6 4 5 The gas inlet port, the buffer space, the first internal space, the second internal space, and the third internal spaceconstitute a space within the sensor element. The space is provided in such a manner that a portion of the spacer layeris hollowed out. The top of the space is defined by the lower surface of the second solid electrolyte layer, the bottom of the space is defined by the upper surface of the first solid electrolyte layer, and sides of the space are defined by side surfaces of the spacer layer.

11 13 30 60 6 10 61 The first diffusion-rate-limiting portion, the second diffusion-rate-limiting portion, and the third diffusion-rate-limiting portionare each provided as two horizontally extending slits (whose openings have a longitudinal direction in a direction perpendicular to the drawing). Furthermore, the fourth diffusion-rate-limiting portionis provided as one horizontally extending slit (whose opening has a longitudinal direction in a direction perpendicular to the drawing), which is formed as a gap with respect to the lower surface of the second solid electrolyte layer. Note that the region extending from the gas inlet portto the third internal spaceis also referred to as a “measurement-object gas flow section”.

43 3 5 43 4 43 Furthermore, a reference gas introduction spaceis provided at a position farther from the front side than is the measurement-object gas flow section. The position is between the upper surface of the third substrate layerand the lower surface of the spacer layer, and a side of the reference gas introduction spaceis defined by a side surface of the first solid electrolyte layer. A reference gas for measuring the concentration of NOx is introduced into the reference gas introduction space. Examples of the reference gas include air.

48 48 43 48 42 An air introduction layeris a layer formed of a porous ceramic material. A reference gas can be introduced into the air introduction layerthrough the reference gas introduction space. Furthermore, the air introduction layeris formed to cover the reference electrode.

42 3 4 48 43 42 42 20 40 61 42 2 The reference electrodeis an electrode formed in such a manner as to be sandwiched between the upper surface of the third substrate layerand the first solid electrolyte layer. As described above, the air introduction layer, which is coupled to the reference gas introduction space, is provided in a vicinity of the reference electrode. Furthermore, as will be described later, the reference electrodecan be used to measure the oxygen concentrations (oxygen partial pressures) of the first internal space, the second internal space, and the third internal space. The reference electrodeis formed as a porous cermet electrode (e.g., cermet electrode containing Pt and ZrO).

10 101 10 11 10 12 11 13 12 11 12 15 13 20 12 101 20 101 10 20 20 11 12 13 20 20 13 21 In the measurement-object gas flow section, the gas inlet portis a portion open to an external space. A measurement-object gas can be drawn into the sensor elementthrough the gas inlet portfrom an external space. The first diffusion-rate-limiting portionis a portion that imparts a predetermined diffusion resistance to the measurement-object gas drawn in through the gas inlet port. The buffer spaceis a space provided to guide the measurement-object gas introduced from the first diffusion-rate-limiting portionto the second diffusion-rate-limiting portion. The buffer spacealso serves as a space (preliminary chamber) for pumping oxygen into the measurement-object gas introduced through the first diffusion-rate-limiting portion. Pumping of oxygen into the buffer spaceis carried out by the operation of a preliminary pump cell. The second diffusion-rate-limiting portionis a portion that imparts a predetermined diffusion resistance to the measurement-object gas, which is introduced into the first internal spacefrom the buffer space. The measurement-object gas is introduced from outside of the sensor elementinto the first internal spaceas follows. Upon pressure fluctuations of a measurement-object gas in an external space (exhaust gas pressure pulsations in the case where the measurement-object gas is an automobile exhaust gas), the measurement-object gas is rapidly drawn into the sensor elementthrough the gas inlet port. Then, the measurement-object gas is not directly introduced into the first internal spacebut introduced into the first internal spaceafter concentration variations of the measurement-object gas are eliminated by the first diffusion-rate-limiting portion, the buffer space, and the second diffusion-rate-limiting portion. As a result, concentration variations of the measurement-object gas, when being introduced into the first internal space, are substantially negligible. The first internal spaceis provided as a space for adjusting the partial pressure of oxygen present in the measurement-object gas, which is introduced through the second diffusion-rate-limiting portion. The oxygen partial pressure is adjusted by the operation of a main pump cell.

15 16 23 6 16 6 12 23 101 16 0 17 16 23 0 16 23 15 12 s s The preliminary pump cellis an electrochemical pump cell including a preliminary pump electrode, an outer pump electrode, and the second solid electrolyte layer, which is sandwiched between the electrodes. The preliminary pump electrodeis provided on substantially the entire surface of a portion of the lower surface of the second solid electrolyte layer, the portion facing the buffer space. The outer pump electrodeis disposed on a portion that is to be exposed to a measurement-object gas that is outside of the sensor element. Of a plurality of electrodes in the measurement-object gas flow section, the preliminary pump electrodeis an electrode disposed most upstream. A pump voltage Vpcan be applied by a variable power supply, which is provided between the preliminary pump electrodeand the outer pump electrode, thereby passing a pump current Ipbetween the preliminary pump electrodeand the outer pump electrode. This enables the preliminary pump cellto pump oxygen from an external space into the buffer space.

21 22 23 6 22 22 6 20 23 6 22 23 a a The main pump cellis an electrochemical pump cell including an inner pump electrode, the outer pump electrode, and the second solid electrolyte layer, which is sandwiched between the electrodes. The inner pump electrodeincludes a ceiling electrode portion, which is provided on substantially the entire surface of a portion of the lower surface of the second solid electrolyte layer, the portion facing the first internal space. The outer pump electrodeis provided on a region of the upper surface of the second solid electrolyte layer, the region corresponding to the ceiling electrode portion. The outer pump electrodeis provided in such a manner as to be exposed to an external space.

22 6 4 20 5 22 6 20 22 4 20 5 20 22 22 a b a b The inner pump electrodeis formed to extend along portions of the upper and lower solid electrolyte layers (second solid electrolyte layerand first solid electrolyte layer), which define the first internal space, and along portions of the spacer layer, which serve as side walls. Specifically, the ceiling electrode portionis formed on a portion of the lower surface of the second solid electrolyte layer, the portion serving as a ceiling surface of the first internal space; a bottom electrode portionis formed on a portion of the upper surface of the first solid electrolyte layer, the portion serving as a bottom surface of the first internal space; side electrode portions (not illustrated) are formed on portions of the side wall surfaces (inner surfaces) of the spacer layer, the portions forming respective side wall portions of the first internal space, the side electrode portions connecting the ceiling electrode portionto the bottom electrode portion; and thus, in the region where the side electrode portions are disposed, the structure has a shape of a tunnel.

22 23 22 2 The inner pump electrodeand the outer pump electrodeare each formed as a porous cermet electrode (e.g., cermet electrode containing Pt and ZrOand containing 1% Au). Note that the inner pump electrode, which comes into contact with a measurement-object gas, is formed by using a material in which a reduction ability that is exhibited to a NOx component present in a measurement-object gas is decreased.

21 20 20 0 22 23 0 22 23 The main pump cellcan pump oxygen from the first internal spaceto an external space and can pump oxygen from an external space into the first internal space. This can be carried out by applying a desired pump voltage Vpbetween the inner pump electrodeand the outer pump electrode, thereby passing a pump current Ipin the positive or negative direction between the inner pump electrodeand the outer pump electrode.

80 20 22 6 5 4 3 42 Furthermore, an electrochemical sensor cell, namely, an oxygen partial pressure detection sensor cellfor controlling the main pump is configured to detect the oxygen concentration (oxygen partial pressure) of the atmosphere in the first internal space. The electrochemical sensor cell includes the inner pump electrode, the second solid electrolyte layer, the spacer layer, the first solid electrolyte layer, the third substrate layer, and the reference electrode.

20 0 80 0 24 0 0 20 The oxygen concentration (oxygen partial pressure) in the first internal spacecan be determined by measuring an electromotive force Vof the oxygen partial pressure detection sensor cellfor controlling the main pump. In addition, the pump voltage Vpof a variable power supplyis feedback-controlled in a manner such that the electromotive force Vbecomes a constant electromotive force, thereby controlling the pump current Ip. As a result, the oxygen concentration in the first internal spacecan be maintained at a predetermined constant value.

30 20 21 40 The third diffusion-rate-limiting portionis a portion that imparts a predetermined diffusion resistance to the measurement-object gas, which has an oxygen concentration (oxygen partial pressure) controlled in the first internal spaceby the operation of the main pump cell, and guides the measurement-object gas to the second internal space.

40 50 40 30 20 40 100 The second internal spaceis provided as a space for further adjusting, by using an auxiliary pump cell, the oxygen partial pressure of the measurement-object gas, which is introduced into the second internal spacethrough the third diffusion-rate-limiting portionafter the oxygen concentration (oxygen partial pressure) is adjusted in advance in the first internal space. With this configuration, the oxygen concentration in the second internal spaceis maintained at a constant concentration precisely; hence, with the gas sensor, the concentration of NOx can be measured accurately.

50 51 23 6 23 23 101 51 51 6 40 a The auxiliary pump cellis an auxiliary electrochemical pump cell including an auxiliary pump electrode, an outer pump electrode, and the second solid electrolyte layer(the outer pump electrodehere is not limited to the outer pump electrodedescribed above, and it is sufficient that the electrode be a suitable electrode positioned outside of the sensor element). The auxiliary pump electrodeincludes a ceiling electrode portion, which is provided on substantially the entire surface of a portion of the lower surface of the second solid electrolyte layer, the portion facing the second internal space.

51 40 51 22 20 51 6 40 51 4 40 5 40 51 51 22 51 a b a b The auxiliary pump electrodeis disposed in the second internal space, and the structure of the auxiliary pump electrodehas a shape of a tunnel similar to that of the inner pump electrode, which is disposed in the first internal spaceas described above. That is, the ceiling electrode portionis formed on a portion of the second solid electrolyte layer, the portion serving as a ceiling surface of the second internal space; a bottom electrode portionis formed on a portion of the first solid electrolyte layer, the portion serving as a bottom surface of the second internal space; side electrode portions (not illustrated) are formed on portions of the respective wall surfaces of the spacer layer, the portions serving as side walls of the second internal space, the side electrode portions coupling the ceiling electrode portionto the bottom electrode portion; and thus, the structure has a shape of a tunnel. Note that, as with the inner pump electrode, the auxiliary pump electrode, too, is formed by using a material in which a reduction ability that is exhibited to a NOx component present in a measurement-object gas is decreased.

50 40 40 1 51 23 The auxiliary pump cellcan pump oxygen present in the atmosphere of the second internal spaceto an external space and can pump oxygen from an external space into the second internal space. This can be carried out by applying a desired voltage Vpbetween the auxiliary pump electrodeand the outer pump electrode.

81 40 51 42 6 5 4 3 Furthermore, an electrochemical sensor cell, namely, an oxygen partial pressure detection sensor cellfor controlling the auxiliary pump is configured to control the oxygen partial pressure of the atmosphere in the second internal space. The electrochemical sensor cell includes the auxiliary pump electrode, the reference electrode, the second solid electrolyte layer, the spacer layer, the first solid electrolyte layer, and the third substrate layer.

50 52 52 1 81 40 Note that the auxiliary pump cellperforms pumping with a variable power supply, the voltage of the power supplybeing controlled based on an electromotive force V, which is detected by the oxygen partial pressure detection sensor cellfor controlling the auxiliary pump. With this configuration, the partial pressure of oxygen present in the atmosphere of the second internal spacecan be controlled to a low partial pressure that has substantially no influence on the measurement of NOx.

1 80 1 80 0 40 30 40 21 50 Furthermore, in addition to this, a pump current Ipis used to control the electromotive force of the oxygen partial pressure detection sensor cellfor controlling the main pump. Specifically, the pump current Ip, which is a control signal, is input into the oxygen partial pressure detection sensor cellfor controlling the main pump, thereby controlling the electromotive force V, and accordingly, the gradient of the partial pressure of oxygen present in the measurement-object gas, which is introduced into the second internal spacethrough the third diffusion-rate-limiting portion, is controlled to be consistently a constant gradient. In the case where the gas sensor is used as a NOx sensor, the oxygen concentration in the second internal spaceis maintained at a constant value of approximately 0.001 ppm by the operation of the main pump celland the auxiliary pump cell.

60 40 50 61 60 61 The fourth diffusion-rate-limiting portionis a portion that imparts a predetermined diffusion resistance to the measurement-object gas, which has an oxygen concentration (oxygen partial pressure) controlled in the second internal spaceby the operation of the auxiliary pump cell, and guides the measurement-object gas to the third internal space. The fourth diffusion-rate-limiting portionhas a function of limiting the amount of NOx flowing into the third internal space.

61 61 60 40 41 61 The third internal spaceis provided as a space for performing a process related to the measurement of the concentration of nitrogen oxide (NOx) present in the measurement-object gas, the measurement-object gas being introduced into the third internal spacethrough the fourth diffusion-rate-limiting portionafter the oxygen concentration (oxygen partial pressure) is adjusted in advance in the second internal space. The measurement of the concentration of NOx is carried out primarily by the operation of a measurement pump cellin the third internal space.

41 61 41 44 23 6 5 4 44 4 61 44 22 44 61 The measurement pump cellmeasures the concentration of NOx present in the measurement-object gas in the third internal space. The measurement pump cellis an electrochemical pump cell including a measurement electrode, the outer pump electrode, the second solid electrolyte layer, the spacer layer, and the first solid electrolyte layer. The measurement electrodeis provided on a portion of the upper surface of the first solid electrolyte layer, the portion facing the third internal space. The measurement electrodeis a porous cermet electrode formed of a material in which a reduction ability that is exhibited to a NOx component present in a measurement-object gas is increased compared with the material of the inner pump electrode. The measurement electrodealso functions as a NOx reduction catalyst that reduces NOx present in the atmosphere of the third internal space.

41 2 44 2 The measurement pump cellcan detect a pump current Ipby pumping out oxygen produced by decomposition of nitrogen oxide in the atmosphere of a vicinity of the measurement electrodeand determining the pump current Ipas the amount of the oxygen produced.

82 44 4 3 44 42 46 2 82 Furthermore, an electrochemical sensor cell, namely, an oxygen partial pressure detection sensor cellfor controlling the measurement pump is configured to detect the oxygen partial pressure of a vicinity of the measurement electrode. The electrochemical sensor cell includes the first solid electrolyte layer, the third substrate layer, the measurement electrode, and the reference electrode. A variable power supplyis controlled based on an electromotive force V, which is detected by the oxygen partial pressure detection sensor cellfor controlling the measurement pump.

40 60 44 61 44 41 2 46 2 82 44 2 41 2 2 After being guided into the second internal space, the measurement-object gas flows through the fourth diffusion-rate-limiting portionin a situation in which the oxygen partial pressure is controlled and reaches the measurement electrode, which is within the third internal space. Nitrogen oxide in the measurement-object gas in a vicinity of the measurement electrodeis reduced (2NO→N+O) to produce oxygen. The produced oxygen is then pumped by the measurement pump cell. At that time, a voltage Vpof the variable power supplyis controlled in a manner such that the electromotive force V, which is detected by the oxygen partial pressure detection sensor cellfor controlling the measurement pump, becomes a constant electromotive force. The amount of oxygen produced in a vicinity of the measurement electrodeis proportional to the concentration of nitrogen oxide present in the measurement-object gas. Accordingly, the concentration of nitrogen oxide present in the measurement-object gas is calculated by using the pump current Ipof the measurement pump cell.

83 6 5 4 3 23 42 83 Furthermore, an electrochemical sensor cell, which includes the second solid electrolyte layer, the spacer layer, the first solid electrolyte layer, the third substrate layer, the outer pump electrode, and the reference electrode, is configured. The partial pressure of oxygen present in a measurement-object gas outside of the sensor can be detected based on an electromotive force Vref, which is obtained by the sensor cell.

100 21 50 41 2 41 In the gas sensor, which is configured as described above, as a result of the operation of the main pump celland the auxiliary pump cell, the measurement-object gas has an oxygen partial pressure consistently maintained at a constant low value (value that has substantially no influence on the measurement of NOx), and, in this state, the measurement-object gas is provided to the measurement pump cell. Accordingly, the concentration of NOx present in the measurement-object gas can be determined based on the pump current Ip, which flows when oxygen that is produced, by reduction of NOx, substantially proportionally to the concentration of NOx in the measurement-object gas, is pumped to the outside by the measurement pump cell.

101 70 101 70 71 72 73 74 75 In addition, to enhance the oxygen ion conductivity of the solid electrolyte, the sensor elementincludes a heater unit, which serves to perform temperature adjustment for heating the sensor elementand maintaining the temperature. The heater unitincludes a heater connector electrode, a heater, a through hole, a heater insulating layer, and a pressure release hole.

71 1 70 71 The heater connector electrodeis an electrode formed in such a manner as to be in contact with the lower surface of the first substrate layer. Power can be supplied to the heater unitfrom outside by connecting the hater connector electrodeto an external power source.

72 2 3 72 71 73 71 101 The heateris an electrical resistor formed in such a manner as to be sandwiched by the second substrate layerand the third substrate layerfrom above and below. The heateris connected to the heater connector electrodevia the through holeand generates heat upon receiving power from outside via the heater connector electrode, thereby heating the solid electrolyte forming the sensor elementand maintaining the temperature.

72 20 61 101 Furthermore, the heateris embedded over an entire area extending from the first internal spaceto the third internal spaceand therefore can adjust the temperature of the sensor elementas a whole to a temperature at which the solid electrolyte becomes active.

74 72 74 2 72 3 72 The heater insulating layeris an insulating layer disposed adjacent to upper and lower surfaces of the heaterand formed of an insulating material, such as alumina. The heater insulating layeris formed to provide electrical insulation between the second substrate layerand the heaterand electrical insulation between the third substrate layerand the heater.

75 3 48 43 75 74 The pressure release holeis a portion provided to extend through the third substrate layerand the air introduction layerand to be in communication with the reference gas introduction space. The pressure release holeis formed to mitigate an increase in internal pressure due to a temperature increase within the heater insulating layer.

90 92 94 90 0 1 2 0 0 1 2 0 80 1 81 2 82 83 0 15 0 21 1 50 2 41 90 17 15 24 21 52 50 46 41 s s The controlleris a microprocessor including a CPU, a memory, and the like. The controllerinputs the electromotive force V, the electromotive force V, the electromotive force V, the electromotive force Vref, the pump current Ip, the pump current Ip, the pump current Ip, and the pump current Ip. The electromotive force Vis detected by the oxygen partial pressure detection sensor cellfor controlling the main pump. The electromotive force Vis detected by the oxygen partial pressure detection sensor cellfor controlling the auxiliary pump. The electromotive force Vis detected by the oxygen partial pressure detection sensor cellfor controlling the measurement pump. The electromotive force Vref is detected by the sensor cell. The pump current Ipis detected by the preliminary pump cell. The pump current Ipis detected by the main pump cell. The pump current Ipis detected by the auxiliary pump cell. The pump current Ipis detected by the measurement pump cell. Furthermore, the controlleroutputs control signals to the variable power supplyof the preliminary pump cell, the variable power supplyof the main pump cell, the variable power supplyof the auxiliary pump cell, and the variable power supplyof the measurement pump cell.

90 0 17 0 15 0 90 0 12 0 12 90 0 0 101 15 20 0 100 0 0 90 15 12 0 0 s s s s s s s s s s s s The controllerfeedback-controls the voltage Vpof the variable power supplyin a manner such that the pump current Ipof the preliminary pump cellreaches a target value Ip*. The controllercontrols the voltage Vpin a manner such that oxygen is pumped into the buffer spaceand does not control the voltage Vpin a manner such that oxygen is pumped from the buffer space. Furthermore, in the present embodiment, for the controller, the target value Ip* is set to a constant value. The target value Ip* is set to a value such that, even in a case where a measurement-object gas outside of the sensor elementis a low-oxygen atmosphere (e.g., atmosphere having an oxygen concentration less than or equal to 0.1 vol %, less than 0.2 vol %, less than 1 vol %, or the like), the measurement-object gas after oxygen is pumped to the inside by the preliminary pump cell(i.e., the measurement-object gas to be introduced into the first internal space) is not a low-oxygen atmosphere. It is to be noted that, in a case where the air-fuel ratio of the measurement-object gas is lower than the stoichiometric air-fuel ratio, that is, the measurement-object gas is a rich atmosphere, unburned fuel is included in the measurement-object gas, and therefore, the oxygen concentration can be determined by the amount of oxygen necessary to sufficiently combust the fuel. In this case, the oxygen concentration is expressed as a negative value. Accordingly, the target value Ip* is set in the following manner, for example. First, a minimum oxygen concentration of exhaust gases from an internal combustion engine that uses the gas sensoris investigated in advance. The minimum oxygen concentration is a minimum among oxygen concentrations in various operation conditions (in some cases, the oxygen concentration may decrease to a negative value). Subsequently, the target value Ip* is set based on the amount of oxygen necessary to increase the minimum oxygen concentration of the measurement-object gas to an oxygen concentration higher than the oxygen concentration of a low-oxygen atmosphere (e.g. oxygen concentration greater than 0.1 vol %, greater than or equal to 0.2 vol %, greater than or equal to 1 vol %, or the like). Since the target value Ip* is set to a constant value, the controllercontrols the preliminary pump cellin a manner such that a constant flow rate of oxygen is pumped into the buffer space. The value of the target value Ip* may be appropriately set based on an experiment as described above. For example, the target value Ip* may be 0.5 mA or greater and 3 mA or less.

90 0 24 0 0 20 0 15 The controllerfeedback-controls the pump voltage Vpof the variable power supplyin a manner such that the electromotive force Vreaches a target value (referred to as a “target value V*”) (i.e., in a manner such that the oxygen concentration in the first internal spacebecomes a constant target concentration). Accordingly, the pump current Ipchanges in accordance with the concentration of oxygen present in the measurement-object gas and the flow rate of oxygen being pumped to the inside by the preliminary pump cell.

90 1 52 1 1 40 90 0 0 1 1 1 1 40 30 40 Furthermore, the controllerfeedback-controls the voltage Vpof the variable power supplyin a manner such that the electromotive force Vreaches a constant value (referred to as “target value V*”) (i.e., in a manner such that the oxygen concentration in the second internal spacebecomes a predetermined low-oxygen concentration that has substantially no influence on the measurement of NOx). In addition, the controllersets (feedback-controls) the target value V* of the electromotive force Vbased on the pump current Ipin a manner such that the pump current Ip, which flows due to the voltage Vp, reaches a constant value (referred to as a “target value Ip*”). As a result, the gradient of the partial pressure of oxygen present in the measurement-object gas to be introduced into the second internal spacethrough the third diffusion-rate-limiting portionis consistently a constant gradient. Furthermore, the partial pressure of oxygen present in the atmosphere of the second internal spaceis controlled to a low partial pressure that has substantially no influence on the measurement of NOx.

90 2 46 2 2 61 61 61 90 2 2 2 61 Further, the controllerfeedback-controls the voltage Vpof the variable power supplyin a manner such that the electromotive force Vreaches a constant value (referred to as a “target value V*”) (i.e., in a manner such that the oxygen concentration in the third internal spacebecomes a predetermined low concentration). Accordingly, oxygen is pumped from the third internal spacein a manner such that the amount of oxygen, which is produced when NOx in the measurement-object gas is reduced in the third internal space, becomes substantially zero. Subsequently, the controllerobtains the pump current Ipand, based on the pump current Ip, calculates the concentration of NOx present in the measurement-object gas. The pump current Ipis a detection value corresponding to the oxygen produced in the third internal space, the oxygen being oxygen derived from a specific gas (in this case, NOx).

2 94 A relationship formula representing a relationship between the pump current Ipand the concentration of NOx is stored in the memory. The relationship formula may be, for example, a linear function formula. The relationship formula can be determined by experimentation in advance.

100 92 90 92 15 21 41 50 0 1 2 80 83 10 11 12 12 15 20 20 40 21 50 61 2 94 92 An example of usage of the gas sensor, which is configured as described above, will be described below. Assume that the CPUof the controlleris in a state in which the CPUis controlling the above-described pump cells,,, andand obtaining the voltages V, V, V, and Vref from the above-described sensor cellsto. In this state, when the measurement-object gas is introduced through the gas inlet port, the measurement-object gas, first, passes through the first diffusion-rate-limiting portionand is then introduced into the buffer space, and, in the buffer space, oxygen is pumped into the measurement-object gas by the preliminary pump cell. Thereafter, the measurement-object gas, which contains the pumped oxygen, reaches the first internal space. Next, in the first internal spaceand the second internal space, the oxygen concentration of the measurement-object gas is adjusted by the main pump celland the auxiliary pump cell, and the adjusted measurement-object gas reaches the third internal space. Subsequently, based on the obtained pump current Ipand the relationship formula stored in the memory, the CPUdetects the concentration of NOx present in the measurement-object gas.

12 15 20 0 2 0 10 s Thus, oxygen is pumped into the buffer spaceby the preliminary pump cellfor the purpose of, as described above, suppressing the measurement-object gas from being introduced into the first internal spacein a state in which the measurement-object gas is a low-oxygen atmosphere. Reasons for doing this will be described. The present inventors investigated pump currents Ipand pump currents Ipobtained by varying the value of the oxygen concentration of the measurement-object gas and the value of the target value Ip*. The measurement-object gas was a gas before being introduced into the gas inlet port. The measurement-object gas used was an adjusted model gas. In the model gas, the base gas was nitrogen, the specific gas component was 500 ppm NO, and the fuel gas was 1000 ppm carbon monoxide gas and 1000 ppm ethylene gas. The model gas was adjusted such that the water concentration was 5 vol % and the oxygen concentration was 0.005 to 20 vol %. The temperature of the model gas was 250° C., and the model gas was flowed through a pipe having a diameter of 20 mm at a flow rate of 50 L/min.

3 FIG. 3 FIG. 3 FIG. 3 FIG. 4 FIG. 4 FIG. 3 FIG. 5 FIG. 4 FIG. 5 FIG. 5 FIG. 0 0 0 0 2 101 s s s is a graph illustrating a relationship between the oxygen concentration of a measurement-object gas and the pump current Ip.illustrates cases in which the target value Ip* was 0 mA, the target value Ip* was 1 mA, and the target value Ip* was 2 mA. The left graph ofis an enlarged view of the portion of the right graph ofencircled by a dashed line (in the left graph, the horizontal axis is on a log scale).is a graph illustrating a relationship between the oxygen concentration of a measurement-object gas and the pump current Ip.shows the same cases as those in.is a graph illustrating in enlarged view a region ofcorresponding to oxygen concentrations of 10 vol % or less. In, the horizontal axis is on a log scale. The oxygen concentration on the horizontal axis is the oxygen concentration of the adjusted model gas, that is, the oxygen concentration of a measurement-object gas that is outside of the sensor element. Furthermore, the A/F of the model gas is also shown in parenthesis on the horizontal axis of. The A/F is a value measured by using a MEXA-730λ, which is manufactured by HORIBA, Ltd.

4 5 FIGS.and 2 0 0 0 2 0 15 2 15 2 s s s s As can be seen from, in the case where the oxygen concentration of the model gas was greater than or equal to 1 vol %, the values of pump currents Ipcorresponding to the same oxygen concentration were substantially the same in all of the cases in which the target value Ip* was 0 mA, the target value Ip* was 1 mA, and the target value Ip* was 2 mA. In contrast, in the case where the oxygen concentration of the model gas was less than or equal to 0.1 vol %, a pump current Ipgenerated in a case where the target value Ip* was 0 mA, that is, the preliminary pump celldid not pump oxygen to the inside at all, had a value smaller than a pump current Ipgenerated in a case where the preliminary pump cellpumped oxygen to the inside. That is, the sensitivity of the pump current Ipto the concentration of NOx was decreased.

3 FIG. 3 FIG. 3 FIG. 3 5 FIGS.to 0 0 0 0 0 0 0 0 0 0 0 0 12 10 20 0 0 2 0 0 21 20 20 0 20 0 20 0 s s s s s s s confirms that, even when the values of the oxygen concentrations of the model gases are the same, the greater the target value Ip*, the greater the pump current Ip. However, with regard to the amount of increase in the pump current Ip, which was determined in comparison with the pump current Ipobtained in a case where the target value Ip* was 0 mA, the pump current Ipobtained in a case where the target value Ip* was 2 mA was not equal to twice the pump current Ipobtained in a case where the target value Ip* was 1 mA. That is, the amount of increase in the pump current Ipwas not directly proportional to the target value Ip*. The reason for this is believed to be that, even when the target value Ip* is large, some of the oxygen pumped into the buffer spaceescapes through the gas inlet portto the outside as a result of diffusion, and therefore not all of the pumped oxygen reaches the first internal space. Furthermore,shows that the pump current Iphad a negative value only in cases where Ip* is 0 mA and the concentration of the model gas was less than or equal to 0.1 vol % (cases shown on the left side of, in which the oxygen concentration was 0.005 vol %, 0.01 vol %, or 0.1 vol %). Thus,confirm that the sensitivity of the pump current Ipwas low when the pump current Iphad a negative value. Negative values of the pump current Ipmean that the main pump cellis pumping oxygen into the first internal space(pumping oxygen in a manner such that the oxygen partial pressure in the first internal spacereaches the target value V*), not pumping oxygen from the first internal space. That is, negative values of the pump current Ipmean that the oxygen concentration of the measurement-object gas to be introduced into the first internal spaceis lower than the oxygen concentration represented by the target value V*.

20 100 15 20 0 20 20 15 12 20 20 15 20 3 FIG. 3 5 FIGS.to The results described above demonstrate that a specific gas measurement accuracy decreases in a case where the oxygen concentration of the measurement-object gas to be introduced into the first internal spaceis low. In contrast, in the gas sensorof the present embodiment, the measurement-object gas, after oxygen is supplied by the preliminary pump cell, is introduced into the first internal spaceas described above, and consequently, as illustrated in, the value of Ipcan be increased (i.e., the oxygen concentration of the measurement-object gas to be introduced into the first internal spacecan be increased). As a result, it is unlikely that the measurement-object gas will reach the first internal spacein a state in which the measurement-object gas is a low-oxygen atmosphere, and consequently, a decrease in measurement accuracy that may occur in the case where the measurement-object gas is a low-oxygen atmosphere is suppressed. From the results in, it is believed that, when the preliminary pump cellpumps oxygen into the buffer spacein a manner such that a measurement-object gas having an oxygen concentration less than or equal to 0.1 vol % does not reach the first internal space, that is, the oxygen concentration of the measurement-object gas reaching the first internal spaceis greater than 0.1 vol %, a decrease in measurement accuracy can be suppressed. Furthermore, it is believed that the preliminary pump cellis to be operated in a manner such that the oxygen concentration of the measurement-object gas reaching the first internal spaceis preferably greater than or equal to 0.2 vol % and is more preferably greater than or equal to 1 vol %.

15 20 22 20 61 20 100 0 1 20 40 0 1 20 20 20 100 15 20 20 In the case where the preliminary pump celldoes not pump oxygen to the inside, a measurement accuracy decreases when the measurement-object gas is a low-oxygen atmosphere. The reason for this is unknown, but, for example, may be as follows. One possible reason is that, when a measurement-object gas that is a low-oxygen atmosphere is introduced into the first internal space, the inner pump electrodeacts as a catalyst to cause NOx to be reduced in the first internal spacebefore the measurement-object gas reaches the third internal space. Another possible reason is as follows. In the case where the measurement-object gas is a rich atmosphere, unburned components such as a hydrocarbon (HC) and carbon monoxide exist in the measurement-object gas. NOx can react with the components, and therefore NOx tends to be reduced in the first internal space. For example, in the instance of a gasoline engine, the air-fuel ratio of the measurement-object gas remains at or near the stoichiometric air-fuel ratio in many cases, and therefore the measurement-object gas may be consistently a low-oxygen atmosphere. Even in such a case, the gas sensorof the present embodiment can detect the specific gas concentration accurately. Furthermore, in the present embodiment, the target value V* is feedback-controlled in a manner such that the pump current Ipreaches a constant value. Possibly, this may also be related to a decrease in measurement accuracy that may occur when the measurement-object gas is a low-oxygen atmosphere. For example, in a case where the oxygen concentration of the measurement-object gas to be introduced into the first internal spaceis temporarily decreased, there is a time lag before the second internal spaceis affected by the influence. Accordingly, there is a time lag before the target value V* is changed based on the pump current Ipto an appropriate value, and as a result, a phenomenon in which, temporarily, oxygen in the first internal spaceis pumped to the outside excessively may occur. Thus, it is possible that, in the case where the oxygen concentration in the first internal spacedecreases excessively as a result of the phenomenon, reduction of NOx may occur in the first internal space. In contrast, in the gas sensorof the present embodiment, it is believed that a decrease in measurement accuracy is suppressed for the following reason: since oxygen is supplied by the preliminary pump cell, the oxygen concentration in the first internal spacedoes not decrease to such an extent that NOx is reduced in the first internal space, even when a temporary decrease in the oxygen concentration of the measurement-object gas, as described above, occurs.

100 1 2 0 1 2 10 0 1 2 0 0 1 2 0 3 5 FIGS.to 6 FIG. 7 FIG. s s Furthermore, in the gas sensorof the present embodiment, spike noise in the pump current Ipand the pump current Ip, which occurs when the atmosphere of the measurement-object gas suddenly changes from a rich atmosphere to a lean atmosphere, or vice versa, can be suppressed. The present inventors investigated the behavior of pump currents Ip, Ip, and Ip, which were generated when the measurement-object gas to be introduced into the gas inlet portwas suddenly changed from a rich atmosphere to a lean atmosphere. The measurement-object gases used were adjusted model gases. The following model gases were prepared: a rich-atmosphere gas having an oxygen concentration of 0.05 vol % and a lean-atmosphere gas having an oxygen concentration of 0.65 vol %. When 30 seconds had passed after the rich-atmosphere gas started to flow through a pipe, the rich-atmosphere gas was replaced with the lean-atmosphere gas. The conditions for the model gases except for the oxygen concentration were the same as those for the model gas used in the measurement associated with. Note that the model gas contains fuel gases (1000 ppm carbon monoxide gas and 1000 ppm ethylene gas) as described above, and therefore a model gas having an oxygen concentration of 0.05 vol % is a rich atmosphere.is a graph illustrating temporal changes in the pump currents Ip, Ip, and Ipin a case where the target value Ip* is 0 mA.is a graph illustrating temporal changes in the pump currents Ip, Ip, and Ipin a case where the target value Ip* is 1 mA.

6 7 FIGS.and 7 FIG. 6 FIG. 7 FIG. 6 FIG. 0 0 1 2 1 2 0 15 12 0 100 0 s As can be seen from, in the case where the target value Ip* was 1 mA (), unlike the case of, the pump current Ipdid not decrease to a negative value and consistently had a positive value even in the time period during which the measurement-object gas was a rich atmosphere (elapsed time was 0 to 30 seconds). Furthermore, in, the spike noise in the pump currents Ipand Ip, which occurred when the rich atmosphere was changed to the lean atmosphere, was reduced compared with. This is believed to be because spike noise tends to occur in the pump currents Ipand Ipwhen the pump current Ipchanges from positive to negative or vice versa. For example, in the instance of a gasoline engine, the air-fuel ratio of the measurement-object gas remains at or near the stoichiometric air-fuel ratio in many cases, and therefore, if the preliminary pump celldoes not pump oxygen into the buffer space, the pump current Ipmay frequently change between positive and negative, and consequently spike noise may frequently occur. In the gas sensorof the present embodiment, such changes in the pump current Ipbetween positive and negative can be suppressed from occurring.

1 2 3 4 5 6 12 15 20 21 61 44 42 82 2 92 90 0 92 94 0 92 23 s Here, correspondence relationships between constituent elements of the present embodiment and constituent elements of the present invention will be clarified. The layered body of the present embodiment in which six layers, namely the first substrate layer, the second substrate layer, the third substrate layer, the first solid electrolyte layer, the spacer layer, and the second solid electrolyte layerare layered in this order corresponds to the element body of the present invention. The buffer spacecorresponds to the preliminary chamber. The preliminary pump cellcorresponds to the preliminary pump cell. The first internal spacecorresponds to the oxygen concentration adjustment chamber. The main pump cellcorresponds to the adjustment pump cell. The third internal spacecorresponds to the measurement chamber. The measurement electrodecorresponds to the measurement electrode. The reference electrodecorresponds to the reference electrode. The oxygen partial pressure detection sensor cellfor controlling the measurement pump corresponds to the measurement voltage detection device. The pump current Ipcorresponds to the detection value. The CPUof the controllercorresponds to the specific gas concentration detection device. Furthermore, the pump current Ipcorresponds to the preliminary pump current. The CPUcorresponds to the preliminary pump control device. The memorycorresponds to the storage device. The pump current Ipcorresponds to the adjustment pump current. The CPUcorresponds to the oxygen concentration detection device. The outer pump electrodecorresponds to the measurement-object gas-side electrode.

100 15 21 20 In the gas sensorof the present embodiment described above, since the preliminary pump cellsupplies oxygen to the measurement-object gas before the oxygen concentration is adjusted by the main pump cell, it is unlikely that the measurement-object gas will be introduced into the first internal spacein a state in which the measurement-object gas is a low-oxygen atmosphere, even in the case where the measurement-object gas is a low-oxygen atmosphere before being introduced into the measurement-object gas flow section. Hence, a decrease in measurement accuracy that occurs in a case where the measurement-object gas is a low-oxygen atmosphere is suppressed.

92 15 0 12 s Furthermore, the CPUcontrols the preliminary pump cellin a manner such that a constant preliminary pump current (target value Ip*) flows, and therefore, with a relatively simple control, oxygen can be supplied to a measurement-object gas that is a low-oxygen atmosphere, in the buffer space.

92 94 100 2 100 100 4 5 FIGS.and In addition, the CPUdetects the specific gas concentration by using a common relationship formula stored in the memoryregardless of whether a measurement-object gas outside of the element body is a low-oxygen atmosphere. As described with reference to, in the gas sensorof the present embodiment, the sensitivity of the pump current Iptends not to decrease even in the case where the measurement-object gas is a low-oxygen atmosphere. Hence, the gas sensorcan detect the specific gas concentration accurately without using different relationship formulas for the case in which the measurement-object gas is a low-oxygen atmosphere and for the case in which the measurement-object gas is not a low-oxygen atmosphere. Hence, the gas sensorcan detect the specific gas concentration readily and accurately.

15 12 23 12 42 42 In addition, the preliminary pump cellpumps oxygen into the buffer spacefrom a vicinity of the outer pump electrode. With this configuration, the following is possible. In comparison with, for example, a case in which oxygen is pumped into the buffer spacefrom a vicinity of the reference electrode, a decrease in measurement accuracy that may occur when the potential of the reference electrodechanges as a result of a voltage drop due to the current during pumping is suppressed.

Note that the present invention is in no way limited to the embodiment described above and may be implemented in a variety of embodiments that fall within the technical scope of the present invention.

92 2 2 94 92 101 0 2 2 2 92 0 2 2 92 2 41 94 92 94 94 2 101 92 94 101 5 FIG. 5 FIG. 4 FIG. 4 FIG. s s For example, in the embodiment described above, the CPUdetects the specific gas concentration based on the pump current Ipand a relationship formula representing a relationship between the pump current Ipand the concentration of NOx stored in the memory, but this configuration is non-limiting. For example, the CPUmay detect the specific gas concentration corrected based on the oxygen concentration of a measurement-object gas that is outside of the sensor element. For example, referring to, according to the data in which the target value Ip* is 1 mA or 2 mA, in the case where the oxygen concentration of the measurement-object gas is consistently less than or equal to 5%, the value of the pump current Ipdoes not significantly change even when the oxygen concentration changes, provided that the actual specific gas concentration (real concentration) is uniform (see). On the other hand, in the case where the oxygen concentration of the measurement-object gas may change over a larger range, the pump current Ipmay change with the oxygen concentration, as illustrated in. In a case where the pump current Ipchanges with the oxygen concentration relatively significantly as just described, the CPUmay detect the specific gas concentration with a correction based on the oxygen concentration. This improves a specific gas concentration measurement accuracy. For example, referring to, according to the data in which the target value Ip* is 1 mA or 2 mA, the pump current Iplinearly changes with the oxygen concentration when the specific gas concentration is uniform, and therefore the relationship between the oxygen concentration and the pump current Ipcan be approximated with a linear function. Accordingly, by using the linear function formula (relationship formula for correction), the CPUmay derive a corrected pump current by excluding an influence of the oxygen concentration from the pump current Ip, which is obtained from the measurement pump cell, and, based on the corrected pump current and the relationship formula stored in the memoryin the embodiment described above, the CPUmay detect the specific gas concentration. In this case, the memorymay also store the relationship formula for correction. Alternatively, in place of the relationship formula stored in the memoryin the embodiment described above, a relationship formula in which the relationship formula for correction is taken into account, that is, a relationship formula representing a relationship between the pump current Ip, the specific gas concentration, and the oxygen concentration of a measurement-object gas outside of the sensor elementmay be stored, and, by using this relationship formula, the CPUmay detect a corrected specific gas concentration. As with the relationship formula stored in the memoryin the embodiment described above, with regard to the relationship formula for correction and the relationship formula in which the relationship formula for correction is taken into account, too, the specific gas concentration can be detected accurately by using the formula, which is a common formula, regardless of whether a measurement-object gas outside of the sensor elementis a low-oxygen atmosphere.

92 92 101 0 0 12 15 0 20 0 0 20 92 15 21 92 101 100 92 101 42 23 92 101 100 s s s In the case where the CPUperforms correction as described above, the CPUmay detect the oxygen concentration of the measurement-object gas that is outside of the sensor element. It is to be noted that the constant pump current Ip(i.e., target value Ip*) corresponds to the flow rate of oxygen pumped into the buffer spaceby the preliminary pump cell. Furthermore, the pump current Ipcorresponds to the flow rate of oxygen pumped from the first internal space. Hence, based on the pump current Ip, the pump current Ip, and the target concentration of the oxygen concentration in the first internal space, the CPUcan detect the oxygen concentration of a measurement-object gas that is a gas before the preliminary pump cellpumps oxygen to the inside and the main pump cellpumps oxygen to the outside. That is, the CPUcan detect the oxygen concentration of the measurement-object gas that is outside of the sensor element. Accordingly, the oxygen concentration necessary for the correction can be detected by the gas sensor. In addition, the CPUcan also detect the oxygen concentration of the measurement-object gas that is outside of the sensor elementbased on, for example, a voltage Vref between the reference electrodeand the outer pump electrode. Alternatively, the CPUmay obtain the oxygen concentration of the measurement-object gas that is outside of the sensor elementfrom a device other than the gas sensor, such as a different sensor or the ECU of the engine, and may use the oxygen concentration for the correction.

15 12 23 12 42 12 12 23 0 17 23 s In the embodiment described above, the preliminary pump cellpumps oxygen into the buffer spacefrom a vicinity of the outer pump electrode, but this configuration is non-limiting. For example, oxygen may be pumped into the buffer spacefrom a vicinity of the reference electrode. With this configuration, the following is possible. In comparison with, for example, a case in which oxygen is pumped to the inside from an external measurement-object gas, oxygen can be pumped into the buffer spaceat a low applied voltage because the reference gas (in this case, air) has a higher oxygen concentration than the measurement-object gas. In contrast, in the case where oxygen is pumped into the buffer spacefrom the vicinity of the outer pump electrode, the voltage Vpof the variable power supplyneeds to be relatively high because, in particular, if a vicinity of the outer pump electrodeis a low-oxygen atmosphere, it is necessary to produce oxygen ions by reducing carbon monoxide, water, or the like present in the measurement-object gas.

13 12 20 13 12 20 In the embodiment described above, the second diffusion-rate-limiting portionis present between the buffer spaceand the first internal space, but this configuration is non-limiting. For example, the second diffusion-rate-limiting portionmay be omitted, and the buffer spaceand the first internal spacemay constitute a single space.

100 61 92 61 92 16 12 16 22 In the embodiment described above, the specific gas concentration detected by the gas sensoris the concentration of NOx, but this configuration is non-limiting, and the specific gas concentration may be the concentration of a different oxide. In the case where the specific gas is an oxide, oxygen is produced when the specific gas itself is reduced in the third internal spaceas in the embodiment described above, and accordingly, the CPUcan detect the specific gas concentration by obtaining a detection value corresponding to the oxygen. Furthermore, the specific gas may be a non-oxide gas, such as ammonia. In the case where the specific gas is a non-oxide gas, the specific gas may be converted into an oxide (e.g., in the case of ammonia, converted into NO). When the converted gas is reduced in the third internal space, oxygen is produced, and accordingly, the CPUcan detect the specific gas concentration by obtaining a detection value corresponding to the oxygen. For example, in a case where the preliminary pump electrodecontains a metal having a catalytic function for promoting oxidation of ammonia, the specific gas can be converted into an oxide in the buffer spacevia the catalytic function of the preliminary pump electrode. A similar configuration is possible for the inner pump electrode. Ammonia is converted into an oxide, which is NO, and therefore the measurement of the concentration of ammonia is performed basically by using the same principle as that for the measurement of the concentration of NOx.

92 15 0 92 0 12 16 42 92 0 101 12 92 101 100 92 0 s s s s In the embodiment described above, the CPUcontrols the preliminary pump cellin a manner such that a constant preliminary pump current (target value Ip*) flows, but this configuration is non-limiting. For example, the CPUmay feedback-control the voltage Vpin a manner such that the oxygen concentration in the buffer space, which is detected based on the voltage between the preliminary pump electrodeand the reference electrode, reaches a target value. Alternatively, the CPUmay control the voltage Vpin a manner such that the lower the oxygen concentration of an outside of the sensor element, the greater the amount of oxygen to be pumped into the buffer space. In this case, the CPUmay detect the oxygen concentration of an outside of the sensor elementby using the method described above or obtain the oxygen concentration from a device other than the gas sensor. Furthermore, the CPUmay control the voltage Vpto be a constant voltage.

0 0 0 101 15 20 0 12 15 2 s s 4 5 FIGS.and In the embodiment described above, the target value Ip* is set based on the amount of oxygen necessary to increase the minimum oxygen concentration of the measurement-object gas, which is the minimum among oxygen concentrations in various operation conditions of an internal combustion engine, to an oxygen concentration higher than the oxygen concentration of a low-oxygen atmosphere (e.g. oxygen concentration greater than 0.1 vol %, greater than or equal to 0.2 vol %, greater than or equal to 1 vol %, or the like). However, this configuration is non-limiting. For example, the target value Ip* may be set to a value such that the pump current Ipdoes not decrease to a negative value even in a case where a measurement-object gas having a minimum oxygen concentration, which is the minimum among oxygen concentrations in various operation conditions of an internal combustion engine, is introduced into the measurement-object gas flow section of the sensor element. That is, the preliminary pump cellmay “prevent the measurement-object gas from reaching the first internal spacein a state in which the measurement-object gas is a low-oxygen atmosphere” by ensuring that “the pump current Ipdoes not decrease to a negative value”. In any case, the amount of oxygen to be pumped into the buffer spaceby the preliminary pump cellmay be set by experimentation in a manner such that, in accordance with a range of fluctuations a component of the measurement-object gas may experience, a decrease in measurement accuracy is suppressed within the range of fluctuations (e.g., in a manner such that, as illustrated in, a state in which the sensitivity of the pump current Ipto the concentration of NOx decreases does not easily occur.

101 100 20 40 61 61 201 201 10 11 12 13 20 30 40 6 4 44 4 40 44 45 45 60 45 44 45 44 51 51 44 201 2 44 8 FIG. 8 FIG. 2 3 a In the embodiment described above, the sensor elementof the gas sensorincludes the first internal space, the second internal space, and the third internal space, but this configuration is non-limiting. For example, the third internal spacemay not be included, as in a sensor element, which is illustrated in. In the sensor element, which is a modified example and illustrated in, the gas inlet port, the first diffusion-rate-limiting portion, the buffer space, the second diffusion-rate-limiting portion, the first internal space, the third diffusion-rate-limiting portion, and the second internal spaceare formed adjacent to one another in such a manner as to be in communication with one another in this order, between the lower surface of the second solid electrolyte layerand the upper surface of the first solid electrolyte layer. Furthermore, the measurement electrodeis disposed on the upper surface of the first solid electrolyte layerwithin the second internal space. The measurement electrodeis covered with a fourth diffusion-rate-limiting portion. The fourth diffusion-rate-limiting portionis a film formed of a ceramic porous member containing, for example, alumina (AlO). As with the fourth diffusion-rate-limiting portionof the embodiment described above, the fourth diffusion-rate-limiting portionserves to limit the amount of NOx flowing into the measurement electrode. Furthermore, the fourth diffusion-rate-limiting portionfunctions as a protective film for the measurement electrode. A ceiling electrode portionof the auxiliary pump electrodeis formed to extend to a position immediately above the measurement electrode. The sensor element, configured as described above, can also detect the concentration of NOx based on, for example, the pump current Ipas with the embodiment described above. In this case, a vicinity of the measurement electrodefunctions as a measurement chamber.

23 15 21 50 41 23 In the embodiment described above, the outer pump electrodeserves as the following electrodes: the measurement-object gas-side electrode (outer preliminary pump electrode) of the preliminary pump cell, an outer main pump electrode of the main pump cell, an outer auxiliary pump electrode of the auxiliary pump cell, and an outer measurement electrode of the measurement pump cell. However, this configuration is non-limiting. One or more of the outer preliminary pump electrode, the outer main pump electrode, the outer auxiliary pump electrode, and the outer measurement electrode may be an additional electrode, other than the outer pump electrode. The additional electrode may be provided outside of the element body and be in contact with a measurement-object gas.

101 1 6 101 1 5 6 101 6 44 6 43 5 4 48 6 5 4 3 42 61 6 1 FIG. 1 FIG. In the embodiment described above, the element body of the sensor elementis a layered body including a plurality of solid electrolyte layers (layersto), but this configuration is non-limiting. It is sufficient that the element body of the sensor elementinclude at least one oxygen-ion-conductive solid electrolyte layer and a measurement-object gas flow section be provided in the interior. For example, referring to, each of the layersto, other than the second solid electrolyte layer, may be a layer formed of a material other than a solid electrolyte (e.g., a layer formed of alumina). In this case, the electrodes to be included in the sensor elementmay be disposed on the second solid electrolyte layer. For example, the measurement electrodeofmay be disposed on the lower surface of the second solid electrolyte layer. Furthermore, the reference gas introduction spacemay be provided in the spacer layerinstead of the first solid electrolyte layer; the air introduction layermay be provided between the second solid electrolyte layerand the spacer layerinstead of being provided between the first solid electrolyte layerand the third substrate layer; and the reference electrodemay be provided behind the third internal spaceand on the lower surface of the second solid electrolyte layer.

90 0 0 1 1 1 90 0 0 0 90 90 0 1 1 1 90 0 80 0 90 0 0 1 In the embodiment described above, the controllersets (feedback-controls) the target value V* of the electromotive force Vbased on the pump current Ipin a manner such that the pump current Ipreaches a target value Ip*, and the controllerfeedback-controls the pump voltage Vpin a manner such that the electromotive force Vreaches the target value V*. However, the controllermay perform a different control. For example, the controllermay feedback-control the pump voltage Vpbased on the pump current Ipin a manner such that the pump current Ipreaches a target value Ip*. That is, the controllermay not obtain the electromotive force Vfrom the oxygen partial pressure detection sensor cellfor controlling the main pump or set the target value V*; the controllermay directly control the pump voltage Vp(therefore, control the pump current Ip) based on the pump current Ip.

22 22 22 51 16 23 42 44 16 22 23 42 44 51 16 22 23 42 44 51 2 2 In the embodiment described above, the inner pump electrodeis a cermet electrode containing Pt and ZrOand containing 1% Au. However, this configuration is non-limiting. It is sufficient that the inner pump electrodecontain a noble metal having catalytic activity (e.g., at least one of Pt, Rh, Ir, Ru, and Pd) and a noble metal having a catalytic activity suppression ability (e.g., Au). The catalytic activity suppression ability is an ability to suppress the catalytic activity of the noble metal having catalytic activity from being exhibited to the specific gas. Also, as with the inner pump electrode, it is sufficient that the auxiliary pump electrodeand the preliminary pump electrodeeach contain a noble metal having catalytic activity and a noble metal having a catalytic activity suppression ability, which is an ability to suppress the catalytic activity of the noble metal having catalytic activity from being exhibited to the specific gas. It is sufficient that the outer pump electrode, the reference electrode, and the measurement electrodeeach contain a noble metal having catalytic activity as described above. It is preferable that each of the electrodes,,,,, andbe a cermet electrode containing a noble metal and an oxide having oxygen ion conductivity (e.g., ZrO). However, one or more of these electrodes may not be a cermet electrode. It is preferable that each of the electrodes,,,,, andbe a porous member. However, one or more of these electrodes may not be a porous member.

0 12 0 92 0 0 0 0 12 92 0 12 12 92 0 101 12 101 0 12 92 0 15 12 12 20 92 15 s s s s s s s s s s The “minimum oxygen concentration, which is the minimum among oxygen concentrations in various operation conditions of an internal combustion engine”, described above, may be, for example, −11 vol % (value of 11 in terms of air-fuel ratio for gasoline engines). For example, when setting the target value Ip* in the manner described in the above embodiment, the following is possible: in a case where a measurement-object gas having an oxygen concentration of −11 vol % flows into the buffer space, an amount of oxygen is necessary to increase the oxygen concentration of the measurement-object gas to an oxygen concentration higher than that of a low-oxygen atmosphere (the higher oxygen concentration may be greater than 0.1 vol %, preferably greater than or equal to 0.2 vol %, and more preferably greater than or equal to 1 vol %), and the target value Ip* may be set based on the amount of oxygen. Similarly, in the case where “the CPUcontrols the voltage Vpto be a constant voltage”, as described in the above modified example, the target value (constant value) of the voltage Vpmay be set in a manner such that, with a pump current Ipthat flows in a state in which the voltage Vpis controlled at a constant value, the oxygen concentration of −11 vol % of a measurement-object gas can be increased to an oxygen concentration higher than that of a low-oxygen atmosphere in the buffer space. In the case where “the CPUfeedback-controls the voltage Vpin a manner such that the oxygen concentration in the buffer spacereaches a target value”, as described in the above modified example, the target value of the oxygen concentration in the buffer spacemay be a value that is in a state in which the oxygen concentration is higher than that of a low-oxygen atmosphere. In the case where “the CPUcontrols the voltage Vpin a manner such that the lower the oxygen concentration of an outside of the sensor element, the greater the amount of oxygen to be pumped into the buffer space”, as described in the above modified example, the following is possible: the correspondence relationship between the oxygen concentration of an outside of the sensor elementand the target value of the voltage Vpmay be set in advance in a manner such that, in a case where a measurement-object gas having an oxygen concentration of −11 vol % flows into the buffer space, the oxygen concentration of the measurement-object gas can be increased to an oxygen concentration higher than that of a low-oxygen atmosphere, and accordingly, the CPUmay control the voltage Vpbased on the correspondence relationship. Similarly, the preliminary pump cellmay pump oxygen into the buffer spacein a manner such that, even in a case where a measurement-object gas having an oxygen concentration of −11 vol % flows into the buffer space, the measurement-object gas does not reach the first internal spacein a state in which the measurement-object gas is a low-oxygen atmosphere. Furthermore, the CPUmay control the preliminary pump cellin a manner such that oxygen is pumped to the inside as just described.

15 16 16 16 16 12 15 12 12 20 15 12 12 20 15 12 12 20 92 15 0 0 12 92 0 92 0 12 92 0 101 12 9 FIG. s s s s s It is to be noted that, in a case where the measurement-object gas is a low-oxygen atmosphere and the amount of oxygen that is pumped to the inside by the preliminary pump cellis too small, the preliminary pump electrodemay cause the specific gas to be reduced because the preliminary pump electrodecontains a noble metal having catalytic activity. Furthermore, in some cases, the catalytic activity of the preliminary pump electrodeis highest at or near the stoichiometric air-fuel ratio (oxygen concentration is 0 vol %, A/F=14.7). In such cases, the following may occur: the preliminary pump electrodereduces a greater amount of the specific gas when the oxygen concentration of the measurement-object gas flowing into the buffer spaceis at or near the stoichiometric air-fuel ratio than when the oxygen concentration is −11 vol % (see also, which will be described later). Accordingly, it is preferable that the preliminary pump cellpump oxygen into the buffer spacein a manner such that, even in a case where a measurement-object gas having any oxygen concentration that is within a range of −11 vol % or greater and 0.1 vol % or less flows into the buffer space, the measurement-object gas reaching the first internal spacehas an oxygen concentration of greater than 0.1 vol %. It is more preferable that the preliminary pump cellpump oxygen into the buffer spacein a manner such that, even in a case where a measurement-object gas having any oxygen concentration that is within a range of −11 vol % or greater and less than 0.2 vol % flows into the buffer space, the measurement-object gas reaching the first internal spacehas an oxygen concentration of 0.2 vol % or greater. It is even more preferable that the preliminary pump cellpump oxygen into the buffer spacein a manner such that, even in a case where a measurement-object gas having any oxygen concentration that is within a range of −11 vol % or greater and less than 1 vol % flows into the buffer space, the measurement-object gas reaching the first internal spacehas an oxygen concentration of 1 vol % or greater. Furthermore, it is preferable that the CPUcontrol the preliminary pump cellin a manner such that oxygen is pumped to the inside in any of the ways just described. For example, when setting the target value Ip* in the manner described in the above embodiment, the target value Ip* may be set to a value such that, even in a case where a measurement-object gas having any oxygen concentration that is within the range of −11 vol % or greater and 0.1 vol % or less flows into the buffer space, the oxygen concentration of the measurement-object gas can be increased to greater than 0.1 vol %. The same applies to the cases described above in the modified examples: the case in which “the CPUcontrols the voltage Vpto be a constant voltage”, the case in which the CPUfeedback-controls the voltage Vpin a manner such that the oxygen concentration in the buffer spacereaches a target value”, and the case in which “the CPUcontrols the voltage Vpin a manner such that the lower the oxygen concentration of an outside of the sensor element, the greater the amount of oxygen to be pumped into the buffer space”.

2 [Investigation of Pump Current Ipin Highly Rich Atmosphere]

4 5 FIGS.and 4 5 FIGS.and 9 FIG. 2 2 2 0 0 100 s s With regard to, described above, the relationship between the oxygen concentration and the pump current Ipin a range in which the oxygen concentration of the measurement-object gas was greater than 0 vol % was investigated. In addition, the relationship between the A/F ratio of the measurement-object gas and the pump current Ipin a case where the measurement-object gas was an even richer atmosphere (highly rich atmosphere) was investigated. The measurement-object gas used was an adjusted model gas. In the model gas, the base gas was nitrogen, the specific gas component was 500 ppm NO, and the fuel gas (unburned component) was ethylene gas. The model gas had a water concentration of 3 vol % and an oxygen concentration of 0 vol %. Further, the A/F ratio of the model gas was adjusted by changing the concentration of the ethylene gas. The A/F ratio was measured by using a MEXA-730λ, which is manufactured by HORIBA, Ltd. The temperature of the model gas was 250° C., and the model gas was flowed through a pipe having a diameter of 20 mm at a flow rate of 100 L/min. Subsequently, similarly to the cases of, the relationship between the A/F ratio of the measurement-object gas and the pump current Ipin cases where the target value Ip* was 0 mA and the target value Ip* was 1 mA in the gas sensorwas investigated. The result is shown in.

9 FIG. 9 FIG. 2 0 15 2 15 0 2 16 22 0 12 20 16 22 12 20 61 16 22 16 22 0 2 0 20 21 20 2 0 2 s s s s s s 2 As can be seen from, in the case where the measurement-object gas had the stoichiometric air-fuel ratio or was a rich atmosphere, that is, the A/F ratio was less than or equal to 14.7, including a case in which the measurement-object gas was a highly rich atmosphere, the pump current Ipgenerated in the case where the target value Ip* was 0 mA, that is, no oxygen was pumped to the inside by the preliminary pump cellhad a value smaller than that of the pump current Ipgenerated in the case where oxygen was pumped to the inside by the preliminary pump cell. That is, in the case where the target value Ip* was 0 mA, the sensitivity of the pump current Ipto the concentration of NOx was decreased. Reasons for this may be as follows. First, the preliminary pump electrodeis, for example, a cermet electrode containing Pt and ZrOand containing 1% Au, as with the inner pump electrode. Thus, one possible reason is that in the case where the target value Ip* is 0 mA, when a measurement-object gas that is a low-oxygen atmosphere is introduced into the buffer spaceand the first internal space, the preliminary pump electrodeand the inner pump electrodeact as catalysts to cause NOx to be reduced in the buffer spaceand the first internal spacebefore the measurement-object gas reaches the third internal space. Another possible reason is that the preliminary pump electrodeand the inner pump electrodemay act as catalysts to cause a reaction between the ethylene gas and NOx present in the model gas. Note that the preliminary pump electrodeand the inner pump electrodeserve as so-called three-way catalysts for NOx and hydrocarbons and that the catalytic activity of the electrodes is high at or near the stoichiometric air-fuel ratio (A/F=14.7). As such, in the case where the target value Ip* is 0 mA, it is predicted that the sensitivity of the pump current Ipto the NOx concentration will likely be lowest at or near the stoichiometric air-fuel ratio. However, in the case where the target value Ip* is 0 mA, when a measurement-object gas that is a low-oxygen atmosphere reaches the first internal space, the main pump cellpumps oxygen into the first internal space, and therefore, it is believed that the A/F ratio of the measurement-object gas at which the sensitivity of the pump current Ipto the concentration of NOx is likely to be lowest shifts to the rich side by a corresponding amount. Accordingly, although no measurement was made, it is also predicted that, in the case where the target value Ip* is 0 mA, the pump current Ipwill have an even smaller value when the measurement-object gas is a slightly rich atmosphere (A/F ratio is approximately 14.4), as shown by the dash-dot line in.

0 12 15 16 22 0 2 s s 9 FIG. In contrast, in the case where the target value Ip* is 1 mA, the oxygen concentration in the buffer spacecan be increased because the preliminary pump cellpumps oxygen to the inside. As a result, it is unlikely that the preliminary pump electrodewill cause NOx to be reduced or will cause a reaction between NOx and a hydrocarbon. Furthermore, almost no reduction of NOx or reaction between NOx and a hydrocarbon due to the inner pump electrodewill occur. Accordingly, referring to, it is believed that in the case where the target value Ip* is 1 mA, the pump current Ipis suppressed from decreasing even when the measurement-object gas is a low-oxygen atmosphere.

5 FIG. 9 FIG. 5 FIG. 2 0 22 0 2 22 16 s s Note that in the embodiment described above, with regard to, the pump current Ipdecreased in the case where the target value Ip* was 0 mA, and, as stated above, a reason for this is that the inner pump electrodeacted as a catalyst. However, as described above with regard to, in the case where the target value Ip* was 0 mA in, too, it is believed that the pump current Ipdecreased because not only the inner pump electrodebut also the preliminary pump electrodeacted as a catalyst.

[Investigation of Durability of Gas Sensor]

100 0 0 100 0 2 101 100 100 100 1 101 100 2 2 2 2 101 100 0 s s s s 9 FIG. 10 11 FIGS.and The durability of the gas sensorwas investigated in the following manner, for a case in which the target value Ip* was 1 mA and a case in which the target value Ip* was 0 mA. First, three types of model gases were prepared. The A/F ratios of the model gases were 12.6, 14.5, and 16.6. The model gases were obtained by adjusting the ethylene gas concentration of a model gas having a similar composition to that of the model gas used for the above-described measurement regarding. Subsequently, a gas sensorin which the target value Ip* was set to 0 mA was prepared, and in a condition at the time of the start of the test (durability period=0 h), the pump current Ipwas measured for each of the three types of model gases used as the measurement-object gas. Each of the measured values was used as a reference value for a sensitivity change ratio [%]. Next, the sensor elementof the gas sensorwas exposed to exhaust gases from a gasoline engine for 100 hours in a state in which the gas sensorwas operated (state in which the gas sensorwas measuring the concentration of NOx). The gasoline engine was a V-type 8-cylinder engine having a displacement of 4.6 L with the air intake mode being natural aspiration (NA). One of the cylinder banks of the engine was used. An operation of the engine was a cycle operation at the air-fuel ratio λ(A/F ratio within a range of 14.3 to 15.1 and exhaust gas temperature within a range of 400° C. to 800° C.) The sensor elementwas exposed to the resulting exhaust gases. After 100 hours had passed (durability period=100 h), the gas sensorwas taken out, and in a manner similar to that for the measurement at the time of the start of the test, the pump current Ipwas measured for each of the three types of model gases used as the measurement-object gas. The rate of change in the pump current Ip, which is a change from the pump current Ipat the start of the test to the measured pump current Ip, was derived, and this was taken as a sensitivity change ratio [%] at a durability period of 100 h. The exposure of the sensor elementfor 100 hours and the measurement of the sensitivity change ratio were repeated in a similar manner, and thus the sensitivity change ratio was measured until the durability period reached 500 h. Also, for a gas sensorin which the target value Ip* was set to 1 mA, the sensitivity change ratio was measured in a similar manner until the durability period reached 500 h. The results are shown in.

11 FIG. 10 FIG. 15 0 100 15 0 15 100 100 s s As can be seen from, in the case where the preliminary pump cellpumped oxygen to the inside continuously with the target value Ip* being 1 mA, the sensitivity change ratio was maintained at or near 0% for all of the three types of model gases even after elapse of time. That is, almost no decrease was observed in the NOx concentration detection accuracy of the gas sensorafter the durability test. In contrast, as can be seen from, in the case where the preliminary pump celldid not pump oxygen to the inside, with the target value Ip* being 0 mA, the following tendencies were observed: for the model gas having an A/F ratio of 16.6, the sensitivity change ratio was maintained at or near 0% even after elapse of time, but for the model gas having an A/F ratio of 12.6 and the model gas having an A/F ratio of 14.5, which were rich atmospheres, the sensitivity change ratio deviated from 0% with time. In particular, for the model gas having an A/F ratio of 12.6, which was a highly rich atmosphere, the sensitivity change ratio significantly deviated from 0% with time and eventually reached a negative value having a large absolute value. This result confirmed that, in the case where the preliminary pump cellpumped oxygen to the inside, the gas sensorhad higher durability, particularly in the case where the gas sensorwas exposed to a measurement-object gas that was a rich atmosphere.

15 44 44 2 100 15 44 2 2 Reasons for this are unknown but may be as follows, for example. First, the following can be assumed. In the case where the preliminary pump celldoes not pump oxygen to the inside, an unburned component (ethylene, in the above-described test) in exhaust gases is adsorbed onto the measurement electrode, which results in a decrease in active sites of the measurement electrode, and as a result, the sensitivity of the pump current Ipin the gas sensordecreases with time. In contrast, it can be assumed that, in the case where the preliminary pump cellpumps oxygen to the inside, the unburned component can be easily converted into, for example, COand HO by being oxidized by the pumped oxygen, and therefore adsorption of the unburned component onto the measurement electrodedoes not easily occur.